Although traditionally considered 2 separate entities based on clinical symptoms and course, schizophrenia (SZ) and autism spectrum disorders (ASD) are heterogeneous, neurodevelopmental disorders with core deficits in social functions. Little is known, however, about the neuropathology associated with these abnormalities in these diagnoses or about the degree and nature of overlap between them. The aim of the current study is to investigate the neural correlates of 3 social cognitive processes, Theory of Mind (ToM), social judgment, and empathy, using a multimodal approach incorporating functional MRI (fMRI) and event related potential (ERP) data, in relation to the nosological diagnoses of ASD and SZ and their clinical and functional symptoms. Eighty SZ and 80 high-functioning ASD patients, ages 18-30, as well as matched healthy controls (HC) will complete the protocol. Participants' social functions will be characterized with self-report questionnaires, computerized tests, role-play and observational tools. Neuronally, ToM will be evaluated with a 2-player interactive fMRI Domino task, social judgment with an fMRI Trustworthiness task and empathy with an ERP Empathy to Pain task. We will (1) identify group differences based on categorical diagnostic mapping in social functions and social cognition-related neural deficits. We will further relate patients' social functions and symptoms to abnormal activation of specific brain circuits; (2) Identify patient sub-groups based on integrated multi-modal neural abnormalities related to social cognitive processes, independent of symptom-based diagnostic categories. We will first use an innovative data-driven method, joint independent component analysis (jICA), to fuse the data from all imaging tasks and to identify the upper 30% impaired patients (either SZ or ASD) and 30% of patients most similar to HC, based on brain activity during the different social tasks. The clinical characteristics of these subgroups will be further evaluated. Then the jICA most discriminative single or integrated components will be entered into a cluster ICA (cICA) to identify natural patient subgroups independently of formal clinical diagnosis. We hypothesize that while both SZ and ASD groups will show abnormal social functions and related brain activity compared to HC, some findings will overlap while others will be diagnosis specific. Importantly, we anticipate that the non-diagnostic based analyses (aim 2) will demonstrate the ability of brain based classifying procedures to identify neurobiologically-based patient subgroups that are more homogeneous than current symptom based categories with respect to clinical and behavioral characteristics. Such subgroups can then be evaluated with regard to differential treatment response and possible etiologies, such as genetic risk variables. If successful, the current study will support the emerging shift in clinical and research paradigms for ASD and SZ specifically and more generally for psychiatric illnesses, toward using dimensional biological (vs. categorical behavioral) measures to identify meaningful patient groups. This in turn will advance etiologic and treatment research for these illnesses.